The participants in this project use a variety of open source and/or commercial desktop GIS software packages. For at least four of the participating entities (three local units of government and one NGO), the tools of choice are the open source packages Quantum GIS (QGIS) and uDig. One of the townships is fortunate enough to maintain an ArcGIS license, so that township uses ArcMap and ArcCatalog for GIS data product creation. Another makes good use of free tools available from Google, namely the utilities and API's contained within Google Earth and Google Maps.
One common drawback to open source desktop GIS packages at this time is their inability to open more than one map window in their map print composers. If the final map product requires one or map insets in addition to the main map, then more than one instance of the GIS software must be used to create the insets. A graphic layout software package such as Inkscape, QuarkExpress or PhotoShop can then be used to create the final map layout and decorations.
Some of the vector-oriented, open source desktop GIS tools familiar to students of Penn State's MGIS program include uDig and QGIS. These two packages fall squarely within the areas of appropriateness for the “Intermediate” and “Experienced” user personas from the Sherman (2008) and Steiniger and Hay (2009) tables reproduced previously in the Open Source Desktop GIS subsection.
The paper by Steiniger and Hay ( 2009) notes map layout and production as a common weakness in open source desktop GIS tools. While this was true at the time of publication of the paper, the current incarnations of QGIS (2.8 at the time of this edit) have vastly increased the package's map composition, styling, and publication capabilities. Some reviewers have gone so far as to venture that some of the capabilities now exceed those available in the leading industry-standard brand.
Again, if the map layout requirements exceed the capabilities of QGIS' Print Composer module, the map data window can easily be exported as a graphic file (usually as a vector SVG file, but often as raster files), which in turn can be laid out and finished using an open source graphics/layout software such as Inkscape, GIMP, and Paint.NET (or their commercial counterparts). Inkscape and its commercial alternatives appear to be the favorite tools for final touch-up work.
Being an open source, standards-compliant package, QGIS can read, convert, and manipulate a very large variety of vector and raster data formats, including open source datastores such as PostGIS and SpatiaLite.
Another advantage of QGIS is its general popularity in the open source GIS community. This ensures that more support for this product exists than is usual for open source products. There are numerous on-line documents, several on-line forums, tutorials both in web page and video formats, and even one book that has been published. (See Resources subsection.)
A screenshot of a GIS map project on QGIS 1.7 is shown above right.
Soon after the start of the project, it became apparent that a serious problem currently exists in using most open source desktop GIS packages here in our state of Michigan. This relates to the use of the Michigan GeoRef map projection by the state GIS data clearinghouse (Michigan Center for Geographic Information) for its publicly-available data layers. The open source GDAL/OGR utilities (see section below) apparently have some issues in correctly handling this projection, which is an instance of the class of map projections known as the Hotine Oblique Mercator. This problem and the use of uDig as part of the workaround is described in the Project Tutorials subsection.
In short, uDig is used in our project's desktop GIS component as a utility program to convert vector data layers from the Michigan GeoRef to the State Plane Coordinate projection systems, not as a desktop GIS map composition package.
uDig also plays an important role in the workflow for creating an open web map that uses the open source GeoServer software for serving maps and other files. See the Open Web Mapping section.
The figure at left shows a typical uDig screen.
MapWindow. The development of this package is carried out by a global team spearheaded by Idaho State University, which has also used the product as a front end for hydrologic data analysis tools. This package is developed using Microsoft's .NET framework, and hence only runs on the Windows platform. QGIS and uDig maintain versions for Windows, Linux, and MacOS. If cross-platform compatibility is not an issue, and Windows is your operating system, take a close look at MapWindow as well. (The figure at left is from the MapWindow website.)
Mapwindow may be of special interest if you are into hydrology and related hydrosciences and engineering. The HydroDesktop software used with MapWindow provides you access to, visualizations of, and analyses numerous publicly-available hydrological and meteorological databases. Check it out!
There are other open source software packages that usefully supplement the desktop GIS packages above. Foremost among them are the command line utilities issued by the Geospatial Data Abstraction Library (GDAL) project. The GDAL Utilities and the OGR2OGR Utility perform numerous format conversion, data transformation, and other tasks on raster and vector files, respectively. The only drawback appears to be their apparent inability to correctly handle Michigan GeoRef projection (or, to be exact, the Michigan Dept of Natural Resources implementation of the Hotine Oblique Mercator/Rectified Skew Orthomorphic projection). This also has adverse consequences for areas of the world that depend on RSO projections for their map work, an example being portions of Malaysia.
For advanced geospatial geostatistical analyses on the desktop, two open source packages stand out.
One is GRASS GIS, which originally started out as a US Army Corps of Engineers product. It was subsequently released to the public domain, and today the center of development for GRASS GIS is in Europe. QGIS has a built-in GRASS plug-in that allows users to access GRASS modules. There are also several open source products based on uDig code that provide a desktop interface to GRASS. An example is jGrass.
A geospatial/geostatistical desktop analysis package with a more user-friendly graphical interface is SAGA GIS, whose development is spearheaded by a team from the University of Hamburg. Currently, QGIS plug-ins are under development to tap into SAGA analysis modules.
Many of these tools also make use of the modules and plug-ins available from the open source R statistical package. Numerous plug-ins for QGIS, GRASS, SAGA, and others make calls to R routines for analyses.
These and other similar packages provide the functionality which, in the ESRI GIS ecosystem, is provided through extensions such as Spatial Analyst and Geostatistical Analyst at additional cost over the license fees for the main GIS desktop product.